[01:01] ok i reported the bug about to chromium [03:46] Pendulum I would like to have teamspeak for accessibility in the chat rooms here on ubuntu .it should make it more accessible for people with some handicaps [03:53] hajour: teamspeak? [03:53] o yes [03:53] I don't know what that is [03:54] a program that make possible to speak instead of typing [03:55] in games they use it to [03:55] leoquant by example have problems with his hands [03:55] well, IRC is something that people use all sorts of methods to access [03:56] it should make it more accessible for him [03:56] it's also proprietary, not open source [03:56] and not everyone can hear [03:57] yes but we have tryed to install a sort of teamspeak in the speechcontrol chat [03:57] *nods* [03:57] others can type [03:57] right, but don't they need to be able to hear what he's saying? [03:57] its just when he neead to say something ore ask [03:58] its no problem every pc have sound [03:58] laptop [03:58] right, but if they can't hear (like are deaf or are in a place where they can't have sound or headphones) [03:58] its not that you need a exspensive soundcard of something [03:58] I understand that. I'm saying we have people who work on our project who are deaf [03:59] someone can say it if it is emportant type [03:59] anyway, to get it into ubuntu channels, someone would need to approach the IRC council [03:59] and I'm not sure they'd agree with it [03:59] i dont mean everybody needs to speak true it [03:59] only the one with a handicap who need it [04:00] btw, I don't know leoquant, but has he tried dasher? [04:00] it's an alternate keyboard input [04:00] mm i shall ask it [04:00] (all mouse, no typing) [04:00] we tryed teamspeak and mumble [04:01] I have hand issues and it's the best thing I've found that worked with linux (it's not voice recognition software, though) [04:01] i just try to find a solution for him [04:01] *nods* [04:01] what does he do for things that aren't IRC? [04:01] hehe Pendulum thats where we working on [04:04] he is working on a beginners team /classroom for the dutch loco [04:04] it is called mwanzo [04:05] hajour: right, but how does he use his computer for non-IRC things? [04:05] one key at a time, only can use one finger [04:05] he should really try dasher [04:08] what would be really nice if there was some sort of conference app with text, sound and video, as easy as irc, cheese [04:08] ok i shall ask him ore he tryed that [04:09] he is in our team for translation [04:09] he have done a lot of work for translation [04:10] cool [04:11] yes and he have al kind of contacts true the translation team in ubuntu-nl (dutch) [04:12] i just want he have a good chance to be active in the speechcontrol to [04:13] and we building a accessibility project [04:14] so it should be weird if there was no accessibility in our team [04:14] I would really suggest he look at dasher at least as a short-term solution [04:14] I <3 dasher and am happy to help with some of the non-obvious bits [04:14] we gonna try that,if it not works i come back here :P [04:15] heh [04:15] well, I will not have the time or energy to take it to the IRC council, but that's how it would have to proceed to get teamspeak in an Ubuntu channel [04:16] Pendulum, we need buttons on programs and that sort of things for applications with contrasting colors for people who are colorblind [04:18] i have asked in design but i for advice but i had not get respons [04:18] sorry for my lines [04:18] well, some of that is not something Ubuntu can control. Some of that is individual programs [04:19] little tired [04:19] otherwise, there are options in compiz to change contrast [04:19] AlanBell I think has played around with that more than I have [04:19] yes but i could not help mhall119 true that [04:19] charlie-tca probably has as well [04:20] AlanBell, and charlie-tca dit also not now a solution for the thing from mhall119 [04:20] (I don't disagree that it should be easier) [04:20] ok [04:21] I don't know if the design team will have any better ideas. and unless we're already talking about on Unity, I'm a little reluctant to look at how the regular gnome desktop looks when Unity is about to be the default for everyone [04:21] orca dit crashed UndiFineD his pc because it couched a log out spontaneously.also after the running [04:22] many times [04:22] we have report the bug [04:23] *nods* I saw that you filed a bug about that [04:23] also the sound from orca cant be used on all pc s audio it seems like it anyway [04:23] I don't know enough about orca and how it works to know if that's an Ubuntu-specific bug or an upstream bug [04:25] i dont no aether i only report :) [04:25] what's the bug number? [04:25] :) [04:25] sec i going to look brb [04:33] UndiFineD, is going to look because i cant find it.but he just is logged out again .stil the results of orca [04:33] so he have to start his pc again [04:34] ok this is him https://bugzilla.gnome.org/show_bug.cgi?id=637621 [04:34] Gnome bug 637621 in speech "Orca crashes, logs me out" [Major,Needinfo] [04:35] oh, it looks like joanie is looking at stuff! that's useful :) [04:38] yes to solve it hopefully [04:38] yeah [04:38] joanie's someone I really like and respect [04:38] so hopefully she'll figure it out [04:39] because its no fun for him to restart his pc at least 10 times on 1 evening [04:40] yeah [04:42] Pendulum, its 5.45 in morning here.i going to sleep now.if you know something ore need me just ping me ore pm.:) [04:43] I'm heading to bed shortly myself. good night! [04:44] o and thanks Pendulum [04:44] good night :) === JanC_ is now known as JanC [16:58] anybody here ever use Sphinx4, it is a voice recognition program [16:59] got a user in #ubuntu who is having some issues with it. [17:00] IdleOne: yes [17:00] AlanBell: you think you can take a stab at it if you got the time? [17:00] user is sherbieny [17:00] ok [17:01] he is getting the follwing error when running from terminal [17:01] gnome-sound-recorder:12900): GStreamer-CRITICAL **: gst_implements_interface_cast: assertion `gst_element_implements_interface (GST_ELEMENT (from), iface_type)' failed [17:03] sherbieny: hi [17:03] hey [17:03] how r u [17:03] #ubuntu makes my head spin, much quieter here! [17:03] I have to do some real life stuff feeding kids in a sec, will have more time later [17:03] yes its like a reading competition there [17:03] sherbieny: good luck with sphinx, hopefully you and AlanBell can get it sorted. [17:04] how did you install sphinx? [17:04] and what version of ubuntu etc? [17:04] I found the accuracy pretty low [17:05] IdleOne: thanks for helping I'll come back to you if I didn't sort my mic issue out [17:07] AlanBell: Well I installed sphinx4 by downloading the bin and src and extracting them together into sphinx4-1.0beta5 [17:07] AlanBell: My specs is: [17:08] AlanBell: Ubuntu10.10 with windows 7 on Hp Pavillion dv6000 [17:11] AlanBell: when I enter: cat /proc/asound/card0/codec#* | grep Codec [17:11] Codec: Realtek ALC268 Codec: Motorola Si3054 [17:13] AlanBell: now My mic is working but not very well first when I run it using the terminal I get this: [17:14] (gnome-sound-recorder:13954): GStreamer-CRITICAL **: gst_implements_interface_cast: assertion `gst_element_implements_interface (GST_ELEMENT (from), iface_type)' failed [17:16] AlanBell: I don't know the problem I seems to have two internal mics [17:17] still there? [17:18] sherbieny, that is a bug to report [17:18] how to report it [17:18] however i get the feeling some of these sound bugs are smp related (more than one cpu) [17:18] it is a gstreamer bug [17:19] I'm trying to understand but I don't [17:20] The gnoem-sound-recorder is recording [17:21] but when I push record button it doesn't start the timer [17:22] AlanBell: A question about sphinx4 [17:23] https://bugzilla.gnome.org/enter_bug.cgi?product=GStreamer [17:26] back now [17:27] AlanBell: How to start a tester or something with WSJ models so that sphinx give me sentences, I read it and sphinx takes what I said and sends me a error percentage report [17:27] hmm, I only have sphinx2 installed at the moment [17:27] AlanBell: and sphinx2 can do what just told you [17:27] AlanBell: and sphinx2 can do what I just told you [17:31] I got one of the WSJ models working with pocketsphinx I think [17:31] so you want it to score your pronounciation or something? [17:33] http://www.keithv.com/software/sphinx/ this kind of stuff [17:35] AlanBell: brb [17:37] Pendulum: http://www.keithv.com/software/speechdasher/ [17:37] http://www.youtube.com/watch?v=Fa7tVUS7yY0&feature=player_embedded [17:37] interesting [17:39] I'm back [17:40] sherbieny: I will try and set up sphinx4 later to reproduce the error you were getting [17:41] there is also a #sphinx channel on freenode [17:41] AlanBell: yes I want some score or some data and also I want to repeat the experiment but changing frequency or something [17:41] AlanBell: no no no no no [17:41] I am quite interested in what you are trying to do [17:42] AlanBell: sphinx4 has no problem at all [17:42] ok, sorry [17:42] AlanBell: I installed it and integrated it with eclipse [17:43] AlanBell: the problem is I can't figure out how to use it [17:43] AlanBell: I can only use the demos like HelloWorld [17:44] AlanBell: but to test the speech engine's accuracy that is the issue [17:45] pretty low! [17:45] but yes, an objective measure would be good [17:45] AlanBell: ofcourse it would very lovely [17:47] I think that is what the NIST tools do, I have seen several references to them [17:47] AlanBell: I need to get over sphinx to find the next speech engine and test it so that I can start documenting my experiments [17:47] there are some evaluation tools here http://www.itl.nist.gov/iad/mig//tools/ [17:47] AlanBell: NIST? is it inside sphinx2 or 4 or is it something else [17:48] I have no idea how to use them though [17:48] it is mentioned in http://www.speech.cs.cmu.edu/sphinx/tutorial.html under word alignment [17:49] we have been looking into speech in general over the past week or so [17:49] it is not yet as awesome as we would like it to be! [17:50] one of the most promising looking things is the simon-listens application which uses the Julius speech recognition engine [17:50] no free software is awesome [17:50] all free software is awesome [17:51] I mean in speech-recognition [17:51] but some need additional awesomeness [17:51] all the softwares I found is so complicated and has no GUIs [17:51] simon-listens is a gui [17:52] and its out!! [17:52] but yes, the recognition engines are toolkits rather than finished goods [17:53] simon-listens is more of a command recognition tool rather than continuous speech recognition [17:53] which is much more practical at the moment [17:53] and thats the issue no speech recognition tool is complete except the commercial products [17:53] like dragon [17:53] from what I have seen free text dictation seems to be a number of years away [17:56] so what are you looking to do with it in the end? [17:56] thats a good question [17:56] I have seen comparison of dns / sphinx / julius / others, dns is marginally better [17:56] didn't dragon get discontinued years ago? [17:57] well it is still there and sold [17:57] It started as an easy topic for my thesis project [17:57] do not know of their development [17:57] AlanBell: Dragon is still around. I assume you're thinking of ViaVoice which used to work on Linux [17:58] Pendulum: probably [17:58] I have to make a comparison between speech recognition engines [17:58] sherbieny: ok, great where are you doing your degree? [17:59] to do a fair comparison I would play each one the same sound file and do some stats on the results [17:59] hajour, is curious an on her way :) [17:59] hai all [17:59] probably using something like the news or a radio play as the sound file [18:00] hi hajour [18:00] reading back:) [18:00] AlanBell: I'm doing it in german university in egypt [18:00] sherbieny: hajour has been looking into speech and pulling together interested people [18:00] sherbieny: wow, interesting! [18:01] Its my graduation project and I'm running out of time [18:01] sherbieny: simon-listens is a german led project, as is openMary [18:02] i can't get anyone of them to work so I can start get the data I need for my documentation [18:02] other universities with strong research groups in the area are Edinburgh where they develop festival (text to speech) and Carnegie Mellon where Sphinx comes from [18:03] Well they are out of my reach now, I hope to find a way to just use sphinx, I don't want to write programs or scripts [18:03] I just want to start something read a sentence and get an error report [18:04] and repeat with changing some variables like frequency, amplitude [18:06] http://www.itl.nist.gov/iad/mig/tools/index.html [18:06] look at the Speech Recognition Scoring Toolkit [18:06] and I hope I can do this on other softwares, write reports, finish my thesis and graduate :) [18:06] yay \o/ [18:06] sherbieny, we have wayward... he produced sonic and announced it on the debian accessibility mailinglist [18:07] sonic can speed and slow voices and alter pitch [18:07] his real name is Bill Cox [18:07] and bedahr from simons listen [18:08] RainCT from julius [18:08] in our team [18:09] and we still like to hear from TheMuso :P [18:09] so if you ever chance your mind [18:09] UndiFineD, what is wayward [18:09] AlanBell, have you let him see the launchpad from speechcontrol team? [18:10] sherbieny, his nick ,his real name is Bill Cox [18:10] his nick is waywardgeek [18:11] you lost me guys :D [18:11] no your still here :) [18:11] hajour: no, I was answering the questions! [18:12] its ok AlanBell :) [18:12] ohh I thought you'r telling me all these info [18:12] https://launchpad.net/~speechcontrolteam [18:12] sherbieny: hajour has been leading a focus group looking into the state of the art in Free Software speech recognition and text to speech [18:12] hajour, my maaaaan please help [18:12] lol i am a she [18:13] ohh, hajour , my womaaaaaan please help [18:13] sherbieny: I think I will have a play with some of those scoring toolkit things later [18:13] hehe corection UndiFineD his women hehe [18:14] sherbieny, about sonic, which could help you: http://lists.debian.org/debian-accessibility/2010/12/msg00030.html [18:14] right now I don't know how to get a percentage score or anything out of a speech recognition system [18:14] i now it is a lot info [18:14] so who's the woman [18:14] hajour [18:14] me [18:15] AlanBell, thanks for the effort and see you soon [18:15] its to much for him :P [18:15] sherbieny: http://www.keithv.com/pub/baselinewsj/baseline_wsj_recipes.pdf [18:15] hope your problem shall be solved soon [18:16] sherbieny: really interested to hear how you get on with it [18:16] hajour, do you know how to use sphinx4 [18:16] mm again a problem with a voice program [18:16] yes [18:16] AlanBell, thanks [18:16] we shall look on it sherbieny [18:16] AlanBell, had it working, i have not got vedics to work yet [18:17] AlanBell, pm? [18:17] sure hajour [18:17] vedics makes use of sphinx4 [18:17] UndiFineD, no AlanBell got sphinx2 working and not the error reports [18:18] oh yes :) [18:18] I have it installed, but needs config [18:18] which i have not yet looked into [18:18] I have it installed to and integrated with eclipse but no use [18:19] all i can do is some demo applications [18:20] and the manual is impossible [18:20] my at-spi is broken, which i need to fix [18:20] http://paste.ubuntu.com/546044/ [18:21] sherbieny: http://cmusphinx.sourceforge.net/sphinx4/tests/live/README.html [18:22] is this an error log from sphinx4 installation [18:22] sherbieny, sorry but i have to make diner now.else i get angry kids.i think you are in good hands with AlanBell and UndiFineD .:) [18:22] thanks [18:23] np :) [18:25] AlanBell, thanks this is the farthest point I reached in my search and I couldn't work it with WSJ [18:25] ok [18:25] I downloaded the -bin zip file so I don't have that, I will grab the -src version [18:27] the src version contains them but when you change the path in config files it get angry and start shouting errors [18:28] yeah, Java doesn't do small friendly errors [20:39] UndiFineD: is your facebook profile pic you, hajour, and the kids? [20:40] yes [20:40] cool :) [20:41] found you :) [20:41] I could tell [20:45] ? [20:46] a ok [20:46] its a pic. from 2 years back on his side :P [20:47] but its the only one we all are on [20:49] :) [20:50] my oldest daughter is here on ubuntu also now [20:52] but more on ubuntu nl.she find english for now a bit to difficult